F4(a). Data about content restrictions to enforce terms of service

The company should clearly disclose and regularly publish data about the volume and nature of actions taken to restrict content that violates the company’s rules.

Elements:

1. Does the company publish data about the total number of pieces of content restricted for violating the company’s rules?
2. Does the company publish data on the number of pieces of content restricted based on which rule was violated?
3. Does the company publish data on the number of pieces of content it restricted based on the format of content? (e.g. text, image, video, live video)?
4. Does the company publish data on the number of pieces of content it restricted based on the method used to identify the violation?
5. Does the company publish this data at least four times a year?
6. Can the data be exported as a structured data file?

Clearly disclose(s) — The company presents or explains its policies or practices in its public-facing materials in a way that is easy for users to find and understand.

Content restriction — An action the company takes that renders an instance of user-generated content invisible or less visible on the platform or service. This action could involve removing the content entirely or take a less absolute form, such as as hiding it from only certain users (eg inhabitants of some country or people under a certain age), limiting users’ ability to interact with it (eg making it impossible to “like”), adding counterspeech to it (eg corrective information on anti-vaccine posts), or reducing the amount of amplification provided by the platform’s curation systems.

Structured data — “Data that resides in fixed fields within a record or file. Relational databases and spreadsheets are examples of structured data. Although data in XML files are not fixed in location like traditional database records, they are nevertheless structured, because the data are tagged and can be accurately identified.” Conversely, unstructured data is data that “does not reside in fixed locations. The term generally refers to free-form text, which is ubiquitous. Examples are word processing documents, PDF files, e-mail messages, blogs, Web pages and social sites.” Sources: PCMag Encyclopedia: “structured data” http://www.pcmag.com/encyclopedia/term/52162/structured-data

“unstructured data” http://www.pcmag.com/encyclopedia/term/53486/unstructured-data

Terms of service — This document may also be called Terms of Use, Terms and Conditions, etc. The terms of service “often provide the necessary ground rules for how various online services should be used,” as stated by the EFF, and represent a legal agreement between the company and the user. Companies can take action against users and their content based on information in the terms of service. Source: Electronic Frontier Foundation, “Terms of (Ab)use” https://www.eff.org/issues/terms-of-abuse

 Indicator guidance: Companies can and should set clear rules about what types of content are not permitted on their platforms or services. This indicator expects companies to publicly disclose data about the actions they take to restrict or otherwise censor content due to breaches to the company’s rules. Publishing this data is an essential first step to holding companies accountable for enforcing their own rules and for the actions they take to moderate content on their platforms and services.

As of the 2020 RDR Index, no telecommunications company received any credit on this indicator. For telecommunications companies, taking action against content violating their rules could entail restricting the dissemination of SMS or restricting access to a web server run by a subscriber. It could also include blocking its subscribers’ access to a specific site, if the block was done to fulfill the company’s own rules rather than an external request (external requests for blocking are covered in indicators F5, F6, and F7).

Companies should publish data about the aggregate number of pieces of content they restrict, remove, or—in the case of telecommunications companies—content they block or filter, as a result of terms of service violations (Element 1). They should also break out this data by violation (Element 2) and by the method—such as a community flagger program or automation—through which the rules violation was detected (Element 3). Companies should also publish this data at least four times a year (Element 4), in line with the Santa Clara Principles, and in a structured data file (Element 5).

• Company transparency report
• Company community standards enforcement report, community guidelines enforcement report, etc.